USA 
United Kingdom
The question that the many who are bewildered by its sudden emergence most commonly ask about AI is, where is all this going to end?
By which they mean... Is this a good thing or a bad thing? Will it take my job? Will I get a household robot? Will someone be able to generate a video of me committing a crime?
All perfectly valid questions.
Recent informed opinion around tech and investment tells us that the question they should be asking is not where but when. And, in answer to that, the cognoscenti suggest, soon.
Whereas 12 months ago all talk was of the brave new world that would be powered by AI, a world of text to video, domestic and commercial robots, driverless vehicles, workplace transformation and so on, now the talk is when, not if, the bubble will burst. And, by way of supplementary consideration, what happens when it does?
Bubbles burst when companies are over-valued and their products over-rated. Think back to the dot-com boom and bust of 20 years ago, when website owners were multi-millionaires one week, based on the over-excited valuation of their online product offering and considering a new career move the next.
Forever blowing bubbles
Throughout history, there have been multiple cases of so-called boom markets catching out heavily invested shareholders and others. In the 18th century a slave trading company (The South Sea Company) was granted monopoly access to international markets in a bid to reduce Britain’s national debt.
Short story: despite its inflated valuation, ‘trading conditions’ (a war) meant it could not deliver what it promised, thousands of investors were ruined and the economy took a huge hit. The economic fallout became known as the South Sea Bubble.
We’re a long way from that but, as happened with the 2008 economic crash, when market sectors are heavily tied in with financial infrastructure the fall-out can be catastrophic.
Artificial intelligence is not yet propping up the financial system in the same way mortgage lending did, so any AI bubble that bursts will likely have an effect only on investors, but will undoubtedly ripple through the stock market. Those that have borrowed to invest may feel the most pain.
'Team mates' | Amazon touts agentic AI 'co-workers' amid job cuts
Spending on AI data centres between 2024 and 2027 is expected to exceed $1.4tn.
Those that only produce AI engines and not other products may be the most affected.
Companies like Meta, Google, Nvidia, and Microsoft, for instance, may take a bath on it, but are likely to be big enough to take the hit without it damaging their core business.
Investment of time
It’s not only about financial investment either. Much time, effort and organizational development has gone into deploying AI into new work systems.
Nvidia CEO Jensen Huang is encouraging all employees to use AI for as many tasks as possible, going as far as to call managers who limit AI usage within their teams “insane”.
Speaking the day after Nvidia reported record earnings, Jensen responded to a question about managers instructing employees to reduce their AI usage, saying: “My understanding is that Nvidia has some managers who are telling their people to use less AI.
“I want every task that is possible to be automated with AI, to be automated with AI. I promise you, you will have work to do.”
The market value of Nvidia has increased eight-fold, to more than $3tn since 2022.
But if (fingers crossed?) it doesn’t quite implode in the way many are predicting and if it merely causes a few casualties as the market right-sizes, then the widespread adoption of AI will continue unabated, re-shaping organizations and roles within them, creating new roles, new product sectors and new career paths.
What new roles will AI create?
Piers Hudson, Senior Research Director at the Gartner HR practice, says: “By 2030, 60% of HR work tasks will be completed through an intelligent agent or LLM-centric interface.
“HR must anticipate where support demands will rise and assess how much their staff must adapt. This means both responding to emerging needs and upskilling HR talent simultaneously.
“Unfortunately, many operational leaders do not involve HR in AI projects because they don’t see the need, or fear HR will slow them down. It’s imperative that HR clarifies its role in an AI-infused workplace and builds an operating model with the right capabilities for this role. Otherwise, HR risks becoming a smaller function monitoring processes for compliance, rather than being a key amplifier of AI’s benefits.”
Hudson anticipates the creation of a number of new roles and job titles including embedded Chief of Staff Roles, managing the people aspects of tech rollouts to achieve performance improvements. As well as...
Global Business Services & “Productization” Leads: As AI automates work across functions there will be demand for integrated systems that manage these processes together. HR will see staff move into roles in centralized AI enablement teams but also have “productization” leads to pass HR processes to those teams.
Community Managers: HR will bring together communities of expertise via community managers to facilitate collaboration, share best practices, and capture valuable insights from these groups.
Work or Worker-Type Specialists: Gathering expertise either in types of specific work (i.e. programming, customer service) or key types of workers – potentially understanding the full lifecycle from hire-to-retire and acting as their talent agents for the organization.
HR “Labs”: HR lab teams will continually find, manage and engineer new sources of data to feed into simulations and scenario-based analyses.
Consensus Facilitators: Expert mediators creating frameworks for reward and well-being that align to organizational strategy.
Product Managers: Promoting a product mindset across HR with greater emphasis on the internal customer experience, fewer silos between HR products or services, and sunsetting services that no longer provide value.
Setting clear expectations
G-P’s 2025 research reflects the boardroom appetite for AI solutions, showing that 74% of executives view AI as critical to the success of their company, while only 1% report they are not using AI in any way.
Adoption is still gaining momentum, with the majority (91%) describing their company’s position as “scaling up.” It suggests a year ahead defined not by pilot projects, but by enterprise-level integration.
Laura Maffucci, Head of HR, G-P (Globalisation Partners), said: “It’s crucial for leaders to understand the value AI brings to the business and set clear expectations for employees on the technology’s purpose and its impact on their roles.”
The next phase will also test leaders’ appetite for risk, governance and trust. The G-P findings show that most leaders trust AI to make compliance, marketing and risk management decisions, although only 3% of executives would trust AI to make any decision.
Even so, enthusiasm is evident with the report saying 95% of executives believe AI tools (ChatGPT, Google Gemini, etc.) are more effective than search engines.
That widening gap between ambition and caution will shape 2026 strategies.
The workforce impact will be impossible to ignore. According to the data, “on average, executives report using AI tools for approximately 40% of their work,” and expectations for productivity gains are rising fast.
Some 67% would rather use AI tools and be 50% more productive, even if it means reducing headcount. As teams shift toward automation, the research argues that workers will become more comfortable using AI technologies, and AI products will be advanced enough to act with limited human oversight and involvement.
Where HR wants AI to go
Whilst it’s harder to call where AI will have taken us at the end of 2026, HR leaders know where they want it to go.
Alisa Di Beasi, CHRO at PHINIA, says: “Ultimately, AI should empower creativity, not diminish it. By automating repetitive tasks, employees gain the time and space to innovate, collaborate, and grow. When technology takes the administrative weight, teams gain the space to experiment and solve problems together.
"AI technology may accelerate how work gets done, but inclusion, empathy and accountability will continue to define effective leadership and a healthy workplace culture.”
Orna Morris, Senior Solution Consultant at Zendesk, added: "In 2026, AI in HR will go beyond buzzwords to help organisations deliver simpler, smarter support for their people - whether it’s surfacing answers instantly through self-service, reducing manual case work, or freeing up HR to provide human support where it is most needed. Used responsibly, this technology doesn’t replace the human touch, it protects trust, clarifies complexity and helps teams feel supported and confident as the world of work keeps changing."
Oliver Shaw, CEO, of Orgvue, calls for a more focused approach to AI adoption: "One of the biggest oversights in AI strategy this year has been not asking some basic questions before deploying AI solutions: What activities or tasks will AI impact ? How does that affect the way in which we design jobs, how many roles will we need, which new ones will emerge? But most of all where are our people under or over-utilised? Without these insights, we know AI initiatives and effective reskilling are likely to stall.
"More than half of companies have admitted they made the wrong decisions about AI-driven redundancies. To avoid a repeat of this, leaders must focus on AI as a tool, not a solution to all problems. To lead effective reskilling movements, organisations must go beyond vague plans and set out clear, structured workforce models that analyse roles at the task level. This will allow HR to identify which teams need to be reskilled, why they need it, and which tools are right for them.
"If 2025 was the year people hoped for a ‘moment of truth’ with AI, 2026 will ask, “How can we plan better?” to be sure that our AI investments of 2025 lead to actual returns in 2026."
Here's the good news.
Yann LeCun, chief AI scientist at Meta, believes another wave of AI innovation is coming down the tracks over the next three to five years because of the limitation of current systems.
“If we want eventually to build things like domestic robots and completely autonomous cars, we need systems to understand the real world,” he says.
“We’re not talking about matching the level of humans yet. If we get a system that is as smart as a cat or a rat, that would be a victory.”
Neither animal has a reputation for great organizational skills, and are unlikely to be keen on collaborating together so, as far as it being a threat to humanity goes, when you put it like that, it suddenly doesn’t seem quite so intimidating.
